Hello team! Happy Friday to all. Question about OpenAI's integration. Is it using the latest mode (4)l? Wouldn't it be nice if users could pick between models as their speed/cost/outcome changes radically?